The measurement of components in body samples is a common feature of clinical assessments. Many diagnoses can be made or confirmed, the state or progress of a disease elucidated, or the risk of many conditions can be assessed, from the concentration of a particular analyte in a body sample, or from the profile of concentrations of several components. The increasing number of known correlations between disease states and the concentrations of one or more analytes makes sample analysis for both single and multiple analytes an increasingly valuable tool. There is thus a correspondingly increasing pressure on clinical laboratories to analyse ever larger numbers of samples increasingly quickly. To satisfy this, there is a need for assays which are quicker, higher throughput, simpler and/or more completely automated.
There is a growing demand for assays to be carried out at the “point of care”. A demand that is increasing with the ongoing transfer of chronic diseases to primary care. This is beneficial for the patient, who receives immediate advice and is left with less uncertainty. It has also been shown that “near patient” testing improves patient compliance, adherence to treatment and therapeutic control (Price. (2001) BMJ 332:1285-8). It is also beneficial for the medical practitioner, who can potentially avoid the need for a multiple appointments and can be more certain of her judgements.
Point of care assays are exceptionally demanding in terms of the need for simple manipulation and rapid results. If an assay takes more than a few minutes then much of the advantage of conducting it at the point-of-care is lost. Furthermore, although the staff conducting such assays are likely to be healthcare professionals, they are not analytical specialists and will not have access to multiple instruments. It is therefore necessary that such assays be designed to rely on the minimum of sample handling. For this reason, and for reasons of time, personnel resources and cost, it is often of great advantage for a plurality of different analyte levels to be determined in a single operation using a single sample on a single analytical instrument and a single test device. This avoids the costly and time consuming-need for extended sample manipulations, multiple test devices or multiple analytical instruments.
One particular problem with point-of-care methods is that they can seldom accommodate calibration. This is of special concern in enzyme based assay methods since enzymes by their nature are sensitive to inactivation during storage.
A common way to avoid the problem of unknown enzyme activity is to determine the end-point of the enzyme reaction. The amount of end-product formed will depend solely on the amount of analyte present at the start of the reaction, providing that there is sufficient reagent to convert all analyte into product and sufficient time is provided for full conversion. As long as there is sufficient active enzyme to catalyse the reaction, the activity of that enzyme will determine only the rate of reaction and not the end-point. However, this approach requires either a great excess of enzyme, which increases cost and/or requires a sufficiently long assay time, which increases total assay time. This is of particular concern to point-of-care assays, which typically should be completed in 5-10 minutes.
Of special concern are enzymatic reactions which do not produce a measurable end-point. Such situations include consecutive reactions where the next reaction begins before the preceding reaction has reached an end-point or where parallel reactions occur. Similarly, where one component is reacted in the presence of a blocked second component that would react in the absence of the “block” then an end-point can only be reached if the blocking is “permanent”, such as by covalent reaction. However, many blocking reagents have only a temporary effect and thus the end point of the reaction of a first component cannot be reached if the second component becomes to any extent “unblocked” during the period required for that end-point to be reached. Typical examples are the measurement of lipid components of specific lipoproteins, e.g. cholesterol associated with high-density lipoprotein. Blocking of lipoprotein components is generally a temporary, kinetic, type of block. This makes typical end-point assays difficult or impossible and other methods such as assays based on fixed-point measurements must be used. These cases are rate-dependent and thus require either fully stable reagents (typically dry reagents), or inclusion of calibrators, to allow compensation for long-term reagent decay. Dried reagents have the disadvantage that they are prone to errors upon reconstitution, often require lengthy reconstitution times, suffer from activity losses during the drying process and, in particular in case of in-device dried reagents, add to the cost of production. Calibrators are frequently not compatible with point-of-care assays.
The structure of a lipoprotein typically has the lipid constituent bound tightly into an inaccessible mass with the protein constituent. Enzymatic reactions will thus generally not affect lipids bound tightly into lipoproteins, or will do so at a very slow rate. Reagent mixtures for the enzymatic reaction of lipids from within lipoproteins thus also typically contain reagents, such as surfactants, that help to “unlock” the lipoprotein and expose the lipid constituents to the action of the enzyme(s). “Blocking” of a lipoprotein component is thus often the stabilisation of that component against the action of such surfactants so that the “blocked” component is not made available for enzymatic reaction. However, over time, the enzymes and/or the surfactants in the reaction mixture will generally begin to cause some degradation of even a blocked lipoprotein component. This then causes a catastrophic effect where the action of the reagents begins to open the lipoprotein structure, which then becomes more reactive and thus more open and so forth. Correspondingly, a “blocked” lipid component can be relied upon to take no measurable part in a reaction for a certain length of time (such as a few hundred seconds) before rapidly and increasingly losing its “block” and causing significant interference. Such temporary blocks cannot be used in an end-point reaction because the block will be broken down before the end point can be reached and thus the result will not represent the desired component.
It is also of note that where the component which it is desirable to measure contains a relatively small portion of the total amount of a particular type of lipid (e.g. less than 30% of the total in a typical healthy patient) then it is particularly important that when the remainder is “blocked”, the assay is carried out before this blocking can become significantly undone. This is because a relatively small unblocking of a large component will have a significant effect upon the result when the smaller component is analysed.
In view of the above, current point-of-care instruments cannot use liquid reagents because they cannot include a calibrator and cannot use end-point analysis because the endpoint is unreachable or unmeasurable. The only solution that has been available for manufacturers of such instruments has thus been the use of stabilised reagents, typically in dried form. This however has its own disadvantages. Not only are dried reagents more expensive but their reconstitution is both time consuming and potentially unreliable. Machines using such dried reagents thus tend to have a higher proportion of anomalous results than expected, probably because of the occasional failure of the enzymatic reagent to be fully reconstituted. A method that allowed reactions having no measurable end-point to be used reliably with solution reagents in the absence of calibrators would thus be of considerable value.
The most common clinical samples taken for assay are fluids, particularly blood and urine, since these are relatively easy to take and to manipulate. In blood, it is typically the content of the fluid plasma which is analysed. Some of the most common and clinically important measurements made on blood-derived samples relate to the plasma lipid contents. The predominant lipids present in blood plasma are phospholipids (PL), triglycerides (TG), and cholesterol (CH). Of these TG and CH are of particular diagnostic interest because of their association with cardiovascular disease, which in turn is one of the most prevalent diseases in the developed world.
Lipids are by their very nature water insoluble and in blood are transported in complex with apolipoproteins, which render them soluble. These complexes, the lipoproteins are classified into five groups, based on their size and lipid-to-protein ratio: chylomicrons, very low density lipoproteins (VLDL), intermediate density lipoproteins (IDL), low density lipoproteins (LDL), and high density lipoproteins (HDL). Chylomicrons are basically droplets of fat, and consist of TG to about 90%. Chylomicrons function as vehicles for the transport of dietary lipids from the ileum to adipose tissue and liver and are present in the general circulation for only a short period after a meal.
The four remaining classes of lipoprotein are produced in the liver. Whereas VLDL, IDL, and LDL are responsible for transporting lipids from the liver to the tissues, the fifth class, HDL is engaged in the reverse transport of superfluous lipids from peripheral tissues back to the liver for further hepatobiliary secretion. VLDL and IDL have short half-lives and deliver mainly TG to the tissues. LDL and HDL have longer half-lives and are the major participants in blood cholesterol homeostasis. On average LDL and HDL combined carry about 95% of the cholesterol present in blood, with LDL carrying about 70% and HDL about 25%.
There are five different types of proteins present in lipoproteins: apolipoprotein (Apo) A, B, C, D, and E, and each type may be further subdivided. The apolipoproteins are important for the formation, secretion, and transport of lipoproteins as well as the enzymatic activities working upon the lipoproteins in the peripheral tissues. Apolipoprotein B (ApoB) is the principal protein in VLDL, IDL, and LDL; in LDL, it is the only protein. HDL is devoid of apoB and its principal protein is apo-A1.
High concentrations of TG are associated with different pathophysiological disorders such as diabetes, cardiovascular disease, hyperlipidaemia, hyperglyceridaemia types I and IV, and nephritic syndromes. Low concentrations are found in hepatic infection and malnutrition.
From many epidemiological studies it is a well established fact that the CH associated with chylomicrons, VLDL, IDL, and LDL is a major risk factor for cardiovascular disease (CVD), with increasing concentrations correlating with increased risk of CVD. The CH associated with LDL particles is considered the main risk factor and is by far the largest of these components. CH associated with HDL on the other hand is inversely correlated with risk for cardiovascular disease. The lower the concentration of HDL, the higher the risk for cardiovascular disease. Therefore it is common practice to determine CH associated with LDL and/or HDL, typically along with total CH, to diagnose and predict cardiovascular disease, as well as in formulating the risk of CVD, potentially in combination with other factors.
Two methods are currently used routinely for quantification of CH. Both methods are enzymatic. In the first method is utilised an enzyme chain beginning with cholesterol esterase and cholesterol oxidase to generate a coloured or fluorescent signal by the generation of hydrogen peroxide. The other method substitutes cholesterol dehydrogenase in place of cholesterol oxidase and determines the amount of CH in the sample on the basis of the amount of the produced. NADH or NADPH. These methods rely on the at least partial release of cholesterol from it lipoprotein carrier before the reaction can proceed effectively. Surfactants are generally used for this function and are well known in the art.
HDL associated CH may be determined by separating, either physically or by blocking, this class of lipoprotein from non-HDL lipoproteins. After making the non-HDL unavailable, HDL associated CH is measured using the enzymatic methods of total CH. This reaction cannot, however be run to its end point because the known methods of blocking are temporary and would become undone before the end-point was reached. The exception to this is where the non-HDL can be physically separated but this requires techniques such as centrifugation which are not available to point-or-care instruments.
Originally, and still much used, was precipitation of non-HDL by using one of the following:
(i) heparin/Mg2+ (Hainline A et al (1982) Manual of laboratory operations, lipid and lipoprotein analysis, 2nd ed. Bethesda, Md.: US department of Health and Human Services, 1982:151 pp),
(ii) phosphotungstate-Mg2+ (Lopes-Virella M F et al (1977) Clin Chem 23:882-4),
(iii) Polyethylene glycol (PEG) (Viikari J, (1976) Scan J Clin Lab Invest 35:265-8), and
(iv) dextran sulfate-Mg2+ (Finley et al (1978) Clin Chem 24:931-3).
The precipitated non-HDL is then removed by centrifugation. The latter method is still recommended by the Cholesterol Reference Method Laboratory Network as reference method for measurement of HDL associated CH (Kimberly et al (1999) Clin Chem 45:1803-12).
Other methods used separation by electrophoresis (Conlin D et al (1979) 25:1965-9) or chromatography (Usui et al (2000) Clin Chem 46:63-72).
The above methods are effective, but require lengthy separation steps and a number of laboratory instruments. In order to eliminate the laborious sample pretreatment two different routes have been taken. Point-of-care instruments have been developed that integrate the separation and quantification of HDL into the test device, which may be cassettes or reagent-impregnated strips, such as the Cholestech HDL assay device and method (U.S. Pat. No. 5,213,965).
For the automatic clinical instruments, homogenous methods were developed which did not require a physical separation of non-HDL lipoproteins to measure the HDL associated CH fraction. The non-HDL particles were blocked by different methods and rendered inaccessible to the CH metabolizing enzymes. The most recent development has been highly specific surfactants that selectively dissolve HDL. IN such a situation, it is the reaction mixture for HDL which effectively “blocks” non-HDL by including only surfactants which leave non-HDL lipids in lipoprotein form and thus largely inaccessible to the action of the enzymes.
LDL associated cholesterol is commonly determined computationally using the Friedewald equation (Friedewald W T et al (1972) Clin Chem 18:499-501):LDL=Total CH−(HDL+TG/2)
Although convenient and in most cases sufficiently accurate, this method suffers from well-known limitations, in particular the need for the patient to fast before being bled (fasting depletes blood of chylomicrons) and the requirement for TG levels to be below 4 g/L. Therefore the NIH sponsored National Cholesterol Education Program (NCEP) Adult Treatment Panel III (ATPIII) guidelines have recommended using direct measurement of LDL associated CH rather than computing it from total CH, HDL associated CH, and TG. However, recent reports question any clinical superiority of directly measured LDL levels over computed levels (Mora et al (2009) Clin Chem 55:888-94).
Originally, LDL associated CH was measured using ultracentrifugation (Havel R J et al, J Clin Invest 1955; 34:1345-53). This is still the most used reference method, but evidently requires sample pre-treatment. Homogenous methods were later developed which did not require a physical separation of non-LDL lipoproteins to measure the LDL CH fraction (U.S. Pat. No. 5,888,827, U.S. Pat. No. 5,925,534).
VLDL associated CH was originally measured using ultracentrifugation. This remains a preferred reference method, but during recent years homogenous methods for determining VLDL associated CH have been developed. These include methods from U.S. Pat. No. 6,986,998 and U.S. Pat. No. 7,208,287.
CH associated with IDL (also called “VLDL remnants” or “remnant-like particles”) is commonly determined using ultracentrifugation, high performance liquid chromatography or electrophoresis. Two homogenous methods were recently developed (U.S. Pat. No. 7,272,047 and US 2007/0161068) that use specific surfactants to further the selective enzymatic decomposition of IDL associated cholesterol.
In recent years several reports have suggested that measurement of nonHDL may prevent more cardiac events than measurement of LDL (van Deventer et al Clin Chem (2011) 57:490-501; Sniderman et al (2011) Circ Cardiovasc Outcomes 4:337-45). In particualt nonHDL may be superior at elevated TG levels (Sniderman et al (2010) J Clin Lipidol 4:152-5). NonHDL is currently not directly measured by any known prior assay method but computed as the difference between total CH and cholesterol associated with HDL (nonHDL=Total CH−HDL)”. The present invention, however allows for direct or indirect (calculated) measurement of non-HDL components. In one aspect of the present invention, at least one of the analytes is bound to the group of non-HDL lipoproteins. In particular, non-HDL cholesterol is a highly preferred analyte. Non-HDL cholesterol may be measured, for example, in a lipid panel assay along with total TG and total CH.
For screening purposes a direct assay for nonHDL will have the advantage of reducing assay time and cost compared to running two assays, Total CH and HDL and computing the nonHDL as the difference (Example 10). Thus, in a further aspect, of the present invention, one analyte will be bound to the non-HD lipoproteins (such as non-HDL cholesterol) and will be measured directly (i.e. not by taking the difference between two other measurements). This analyte may be measured with or without any other analytes. Furthermore, the advantages of measuring non-HDL (eg non-HDL CH) directly extend to assays which used end-point estimation (i.e. the calculation of a fictive end point as described herein) and also assays that use conventional techniques. Thus in a further aspect the invention provides an assay for the direct measurement of a lipid component bound to non-HD lipoprotein, such as non-HD cholesterol. A corresponding method for assigning a risk of or propensity to CVD is provided by comparing such a value to an appropriate threshold, such as a threshold derived from populations of healthy individuals and/or individuals suffering from a high risk or propensity to CVD.
TG is determined routinely in a four step enzymatic reaction, in which lipoprotein lipase hydrolyzes TG to unesterified glycerol and free fatty acids. The glycerol is then phosphorylated (glycerokinase) and oxidized (glycerol-3-phosphate oxidase) to di-hydroxy-acetone-phosphate and hydrogen peroxide, which is used to generate a coloured, fluorescent or chemiluminescent signal.
As with the measurement of CH of different lipoprotein classes, measurement of TG of specific lipoprotein classes may be performed by several methods that exploit different chemical and physical characteristics of the different lipoprotein classes.
As discussed above the measurement of a lipid component of a specific lipoprotein class, e.g. cholesterol in HDL, constitutes a particular problem for point-of-care assays. Current methods depend on temporarily blocking the particular lipid component present in lipoproteins other than the specific lipoprotein class and then measuring the particular lipid component associated with the specific lipoprotein class. Such methods rely on blocking the unwanted lipoproteins with synthetic polymers and polyanions (U.S. Pat. No. 5,773,304, U.S. Pat. No. 6,811,994) or antibodies (U.S. Pat. No. 4,828,986, U.S. Pat. No. 6,162,607) or cyclodextrin combined with polyethylene glycol modified enzymes (U.S. Pat. No. 5,691,159), or use specific surfactants (U.S. Pat. No. 7,208,287, U.S. Pat. No. 7,272,047).
Because the blockage is temporary a true end-point for these reactions is not possible to measure, the only true end-point is the end-point of the particular lipid component present in all four classes of lipoprotein. The particular lipid component of the specific lipoprotein is therefore measured either at a fixed time point, chosen so that the particular lipid component present in lipoproteins other than the specific lipoprotein does not interfere substantially (usually 5 minutes), or kinetically during the first minutes of the reaction. In both cases is required either fully stable reagents, i.e. dry reagents, or inclusion of calibrators, to compensate for long-term reagent decay.
A different approach is to eliminate selectively the particular lipid component associated with lipoproteins other than the specific lipoprotein class being analyzed, in an enzymatic reaction not giving detectable product. The particular lipid component of the specific protein class is then converted into a detectable product. Several such methods have been described making use of surfactants that react selectively with specific lipoprotein classes (EP 1164376, U.S. Pat. No. 5,925,534, U.S. Pat. No. 6,194,164, U.S. Pat. No. 6,818,414, U.S. Pat. No. 6,893,832).
However, in order to accomplish complete elimination the reaction must be allowed to reach end-point, and that takes time. The time required for this would be too long time for point-of-care assays and in particular for point-of-care assays measuring a plurality of analytes, (known as “panel” assays). In addition, these approaches have problems with inaccuracies caused by an incomplete elimination of the particular lipid constituent in lipoproteins other than the specific lipoprotein being analyzed or by non-specific elimination of the particular lipid constituent in the specific lipoprotein. In practise, then, these types of assays also require fixed-point measurement and thus rely on either fully stable reagents, i.e. dry reagents, or inclusion of calibrators, to compensate for long-term reagent decay.
In view of the above, there is evidently a need for improved point-of-care assays for measuring one or a plurality of plasma components (such as lipid components) and in particular for assays including lipid components of specific lipoprotein classes. It would be advantageous if such assays allowed for liquid reagents, did not require calibrators and/or had short total assay times.
The inventors now have established that it is possible to construct an assay method, suitable for the point-of-care apparatus that uses blood-cell containing samples (such as whole blood), uses liquid reagents, has a short total assay time and does not require calibrators.
The inventors surprisingly have found that it possible to determine an end-point for an enzymatic reaction where that end point is theoretically unreachable and/or in practice unmeasurable. This allows many of the advantages of end-point analysis to be applied in situations where end-points have previously not been considered. For example, in a sequence involving several enzymes, product will be consumed and thus the end point cannot be reached. Also, in those cases where an end-point cannot be directly measured, such as in consecutive enzyme reactions where the next step starts before the preceding reaction has reached an end-point, or in parallel enzyme reactions where there is a sufficient difference in the early progress of the reactions.
In situations where the end point cannot be or is not measured, it may be computed using a suitable algorithm rather than directly measured, with accuracy and CV similar to those achieved with direct measurement. What is needed is for the reaction to be monitored for a sufficient length of time and then using a suitable algorithm to accurately predict the end-point. Suitable algorithms and equations have been regularly used in the art for curve-fitting purposes and are thus well-known, but have not previously been applied in this way to predict an end point that cannot be measured. Typically, the reaction should be monitored until at least 50% of the target analyte has been converted, although this may constitute only a minor fraction of the progress curve. Monitoring the reaction until at least 40%, preferably at least 50% and optionally at least 60% of the target analyte has been consumed is appropriate in various embodiments.
In one embodiment, it is preferable that the measurement interval is chosen such that any parallel reactions (such as unblocking and reacting of any blocked components) are not significant.
In an alternative embodiment, parallel reactions such as unblocking and reacting of any blocked components may take place to a measureable extent but the influence of such parallel reactions may be to a certain extent corrected for by post-measurement analysis where more than one analyte from the sample is measured. For instance when a plurality of lipid analytes are measured, such as in a lipid panel assay, the interference of nonHDL on HDL measurements (such as by partial unblocking and reaction of the non-HDL component) may be partly corrected for in an iterative process involving the readings for HDL and total cholesterol, according toTotal cholesterol=HDL+nonHDL,and a predetermined standard curve for the effect of nonHDL on the HDL assay (Example 6).
The inventors also have found that using end-point estimation with reactions that produce measurable end-points may afford substantial advantages to point-of-care assays using liquid reagents. Measurement of end-points must take into consideration the effect on assay time of loss of reagent activity with storage time, using estimated end-points avoids this and consequently allows for shorter measurement times to be used.